Autonomous UAVs in Critical Infrastructure Inspection: Empirical Evaluation, Reporting Standard, and Implications for Strategic Decisions
Purpose: Autonomous unmanned aerial vehicles (UAVs) have become a mature tool for remote sensing data acquisition for critical infrastructure (IC) inspections, particularly in the energy, transportation, bridge infrastructure, water networks and industrial installation sectors. The aim of the article is to synthetically capture the results of empirical research: field inspection missions using RGB, thermal imaging (TIR) and LiDAR sensors, as well as automatic data analytics (deep learning) to detect defects and anomalies. Design/Methodology/Approach: In the methodological part, a unified scheme for reporting empirical research (mission parameters, method of obtaining the “reference truth”, quality metrics and operational utility indicators) is proposed, as well as a framework for integrating results with crisis management processes (time-to-assessment, false dispositions, repair priorities). Findings: A review of representative studies indicates that the effectiveness of component and fault detection depends not only on the selection of the machine learning model, but above all on the quality and standardization of data, environmental conditions, and the method of field validation. Practical Implications: Comparable reporting enables robust synthesis across sectors and accelerates transfer to practice. Integrating UAV products with GIS/CMMS/SCADA and tracking time-to-assessment and false dispatch supports crisis-management decisions and repair prioritization. Originality/Value: A minimal, six-block protocol is proposed to standardize reporting of empirical UAV inspections of critical infrastructure.